distribution of bug types (such as null pointers, API misuse, etc.) in defects4j

[regminer]

I would like to know if there is a distribution of bug types (such as null pointers, API misuse, etc.) in defects4j stated.

[rjust]

There are some classifications based on structure and syntax (e.g., here). You can also find various dissections of the data set (see top-level README).

While NPEs are fairly unambiguous and easy to identify (these are explicitly listed in the trigger-test files), API misuse bugs require a more precise definition (and are likely more subjective). If you have definitions and rules for bug-type classifications, we'd be happy to discuss infrastructure support for (semi-)automated classification.